The cloud service provider announcements are coming thick and fast and today it’s the turn of OpenStack SP Metacloud to join the fray. Metacloud deploys and, more importantly, supports OpenStack based clouds. The product is delivered as a service and essentially layers OpenStack on top of organizations” existing hardware. It’s another attempt to eek out some greater efficiencies from hardware by slapping a cloud stack on top of them – that’s not a criticism, sure it’s way sexier to just jump right into the public cloud vendor de jour, but the reality is that for a CIO with significant sunk costs in existing hardware, someone who can come and offer them, as a readily obtainable service, higher levels of efficiency is a pretty welcome vendor. Metacloud was founded in 2011 and includes architects from both Yahoo! and Ticketmaster. The company is an OpenStack foundation member and a contributor to the project.
The interesting thing for me about Metacloud is that they’re unashamedly a service based offering. While many cloud companies try and wrap what they do up in a product skin, the reality with an open source cloud stack like OpenStack is that the differentiator is service It’s an approach long taken by Rackspace although arguable that company, with it’s varying different areas of focus, is less able to deliver upon it than a smaller, more nimble player like Metacloud.
Metacloud’s Carbon OS includes 24×7 support from Metacloud’s in-house cloud team whom monitors, troubleshoots, upgrades, handles capacity planning, and perform bug fixes. While many organizations would consider at least some of these functions as best being delivered from within the organization – if you subscribe to the theory that cloud is, at its most fundamental, about abstracting responsibility for non-core operations away from the customer, then it’s also logical to move responsibility for as much of that process to a third party service provider.
Given the background of the founders (the founders have previous large scale infrastructure engineering experience) it makes sense that they can offer a higher level of skill and service to organizations that otherwise couldn’t afford this caliber of engineer on their ops team. People question the cost impact of a service provider taking responsibility for higher levels of a customer’s functional needs – but the theory goes that:
- The economies of scale mean that the SP can offer a high level of service at a lower cost despite the margin they charge
- In any case most organizations could afford this gilt-edged engineering talent
Of course this opportunity – eking out a few more percentage points in terms of efficiency from existing kit – is eventually going to fall away. This in no means dismisses the value of what Metacloud is doing, if anything it simply questions the potential returns that its venture capital backers can make. That said there is a huge amount of legacy kit out there and a massive number of organizations that aren’t in a position to throw it away in a wholesale move to the public cloud – for these customers Metacloud is a compelling offering.
Well, I suppose there is room for faux clouds that layer on top of existing virtualization environments. And some day those aging compute assets can be cycled-out for a real cloud infrastructure.
No, I don’t “subscribe to the theory that cloud is, at its most fundamental, about abstracting responsibility for non-core operations away from the customer.” If a “cloud” doesn’t pass the sniff test of the NIST definition, it isn’t a cloud.